Date: October 21st, 2008
Article by: Nathan Glentworth (Owner / Head Editor)
Product was submitted by: Sapphire
<--CLICK FOR PRICES ON A SAPPHIRE VIDEOCARD IN AMERICA
<--CLICK FOR PRICES ON A SAPPHIRE VIDEOCARD IN CANADA
As with any videocard, overclocking is an art form that requires patience to find the right balance of performance while keeping the videocard 100% stable. When we boil the overclocking experience of the TOXIC HD4850 down, we are left with a couple key points.
Even though the core is remarkably cool, the core is not happy to be pushed anymore beyond the 675Mhz this card is clocked at. Even 5Mhz more and it will kick out during testing which will require a reboot of the computer. Keep in mind that this core is already overclocked 50Mhz beyonds the RV770's stock 675Mhz core so I wouldn't deem it a failure that it can't go anymore. It has already been pushed to its limit right out of the box which is fine with me.
As for the memory, the overclocking situation is much better with the maximum overclock I could perform with no errors being 1150Mhz which is 50Mhz faster than what you get out of the box and is 150Mhz faster than you standard HD4850. This does impact performance and does yeild a small noteable gain in synthetic benchmarking and gaming performance as you will see in the upcoming benchmarking results.
The Testbed is comprised of:
Intel QX9770 Extreme Quad Core 3.2Ghz Processor Check Prices
Gigabyte P35-DS4 Premium Motherboard Check Prices
Crucial Ballistix Tracer 2Gig PC-8500 DDR2 Memory Check Prices
Windows XP Professional SP3 Check Prices
Maxtor 250gig SATA-150 HD Check Prices
PCI Express Videocards Benchmarked Against:
Gigabyte 512Meg 8800GTX Check Prices
XFX 512Meg 8600GTS XXX Edition Check Prices
Sapphire 512Meg 2600XT Check Prices
Benchmarking Programs Are:
Half Life 2 Episode 2 HOC Benchmark
Far Cry ( 8x AA / 16x AF)
Call of Duty 2 TechPowerUp Benchmark (Maximum Graphics Quality)
Crysis ( 8x AA / 16x AF)
Crysis WarHead ( 8x AA / 16x AF)
All tests are done using DX9c seeing we will be running Windows XP for maximum performance. In most games today, there is almost no difference in visual quality that can noticed when you compare DX9c vs DX10.1. DX10.1 involves installing Vista which in itself is a huge performance hindrence that most gamers avoid. Relative performance comparisons can be made seeing the difference will be relatively the same for each DirectX version.
Throughout the testing of this videocard, I will be showing you how this card performed on the system mentioned above using the following programs at 1024x768 and 1280x1024 resolutions. Higher resolutions will be added over the coming months.
Both testing situations will be benchmarked with stock and overclocked settings initialized separately.